No less than four notes on less.

ABSTRACT

Here are four notes concerned with related aspects of the morphosyntax of less in English. The first is concemed with (re-)drawing attention to the currency of a usage in which less is accompanied by a plural noun, as in the title of the article. The second aims to show that less and its ancestors are heads of a partitive construction, whether this is marked overtly or not. In the third note we are concerned with the explicit characterisation of the status of vatious determiners, including less, with respect to countability and collcctivity. Finally, note four examines the status of less as a comparative, and attempts to resolve thereby two anomalies observed in note 1. I do not address here, for the most part, the "adverbial" uses of less and its ancestors. (1)

1. Five items or less

In its article on less the OED deseribes the usage illustrated by (1):

1) I thinke there are few Vniversities that haue lesse faultes than Oxford, many that haue more. (Lyly 1579 [1923]: 208)

as "now regarded as incorrect". And this seems to be a view shared by many who concern themselves with such things. The general feeling (as embodied in e.g., Fowler 1926: 321; Gowers 1973: 227-228) seems to be that when a plural noun is aceompanied by a paueal quantifier that quantifier should be few, as with the first quantifier plus plural noun combination in (1).

However, we continue as users of English to encounter and perhaps produce instances of less plus plural noun, and not just in such manifestations as the supermarket sign that gives this note its title, where the censorious might be apt to find simply signs of ill-education (and the less censorious might get excited about social variation rather than, or as well as, syntax and semanties). Juul (1972: 1, n. 3, 1975: 215, n. 3), for instance, notes the examples in (2) from his corpus of (then) largely recent texts: (2)

2) a. But there were less casualties than might have been expected ... (Orwell 1938: 84)

b. The metropolitan area itself has a million less residents and half a million fewer jobs than in 1939. (NS: 413)

c. We have less grounds for optimism about the immediate financial outlook. (ST: 63)

d. Sir,--Mr Bilan Allison ... makes a plea for more coherence and less frills in art education. (TES: 2)

e. When we land at New York, this door will lead you to Pan Am's shiny new 93 million dollar terminal. It's the slickest, fastest terminal at Kennedy--less steps to the street than from any other terminal. (Foylibra: 7)

What is striking is that these quantified phrases all share a semantic property: the quantified phrase is to be interpreted collectively.

We can illustrate collective vs. distributive/singulative in relation to a sub-type of those nouns which are often referred to as "collective" nouns--rather unhelpfully so called given that they can be used either collectively or distributively, as in (3a) (collective) vs. (3b) (distributive):

3) a. The committee is/are convening right now.

b. The committee are gnashing their teeth.

"(Intransitive) convening" is a necessarily collective aetivity; "gnashing teeth" is distributed throughout the individual members of the group. The collective use of sueh nouns, what I shall refer to as group nouns, permits but does not necessitate a singular verb concord.

It might be argued that in the uses in (2) the nouns are perhaps even lexicalised as plural and collective expressions of a collection seen as a unity. But even this might give us pause, at least, in simply seeing the use of less with plurals as "incorrect". However, let us confront the argument that the occurrence of less here is indeed simply a reflection of lexicalisation of these plural forms, as Fowler (1926: 321) suggests concerning troops and clothes. In this case, we would have to recognise that these lexically plural nouns also have, in the case of those in (2d, e), at least, regular count congeners, as in a frill, a step; and with all of them fewer is available as a non-synonymous alternative to less. Moreover, colloquially, at least, and given the appropriate circumstances, almost any count noun can be given an appropriate collective interpretation if quantified with less. Consider, for example, the range of possibilities in (4):

4) What we want in government is less women/lawyers/ABs/lords/ do-gooders ...

Again, the use of less here insists on a collective interpretation; (4) is concerned with the size of a grouping. In this case the context almost forces such an interpretation on the otherwise normally distributive fewer, but less is more insistently anti-distributive.

There is one circumstance in which the collectivity of less, compared with fewer, is particularly evident; and it is also one where the association of less with a plural interpretation is especially difficult to dismiss as "incorrect", given its prevalence. Consider (5), where a numeral is the source of the comparison:

5) a. She earned less/?fewer than twenty pounds.

b. She weighed less/*fewer than 100 pounds.

Poutsma (1914: 302-304, [section]16, 1916: 1104, [section]76) cites, but scarcely comments on, such examples. But it is one area of usage that has had its defenders (Jespersen 1949: 380). The interpretation is strongly collective (She earned the sum of less than twenty pounds), and fewer is, indeed, disfavoured--strongly so in (5b) (suggested to me by Graeme Trousdale). The context has to be made more favourable to a distributive interpretafion for fewer to become more acceptable in such circumstances:

6) She received fewer than five replies.

There is also an ambivalence in the structure of (5). Is the noun the complement of twenty or of less? Is the comparative source than twenty or than twenty pounds? It may be that this ambivalence shields this particular construction from proscription. I shall return fo this ambivalence in the final note.

My titular example for this note fits a variant of the pattern illustrated by (5), as an elliptical variant of (7a):

7) a. Five items or less than five (items).

b. Five or less than five items.

(7b) is the (5)-type variant. Where we have anaphora by pronoun rather than ellipsis in the (7a) variant, as in (8), we find a singular pronoun, in accord with the collective interpretation:

8) Five items or less than that.

There is no such equivalent for (7b), however. This too will attract out attention in the final note.

One moral of this little tale, not a novel one, I admit, is that we have to be careful exactly what we are stigmatising as "incorrect". In this particular case, less plus plural noun is more widely attested than simply dubbing it "incorrect" would suggest. The currency of the type of (5) in particular contradicts this. We seem to have a less flamboyant example of the stigmatisation based on premature analysis that is more flagrantly displayed in the history of attitudes to "double negatives" in English. Imputations of "incorrectness" or "illogicality" should be based on what usage actually is and what its own logic might be.

Another, more descriptive aspect of our tale is that there are at least two apparently incidental, and unconnected, observations made here that need to be commented on further: the structural ambivalence illustrated by (5)-(7), and the absence of a pronominal equivalent to (7b). I want to try to show that these are related, and, as indicated, I shall return to both these observations in note 4. But this will involve us in being a little more formal. I introduce the necessary formalism via a partially diachronic look, in the next note, at the constructions introduced in the present note.

2. No laes [eth]onne xx scipa

The OED takes the Old English (OE) overtly partitive construction in (9) to be the ancestor of the stigmatised use of less illustrated by (1) in the previous note:

9) a. [thorn]am [thorn]e bi[eth] laes synna forgifen, se lufa[eth] hwonlicor 'to-him that is less of-sins forgiven, he loves less' (AElfric Homilies: Dedication of a church: 60)

b. for[eth]an [thorn]e her bi[eth] laes manna on wodnes daeg, [eth]onne nu to daeg beo[eth] 'because that here is less of-men on Wednesday than now today are' (AElfric Ash Wednesday: 52)

c. & gif [thorn]aer beon laes manna [thorn]onne [thorn]aet lamb maege fretan ... 'and if there be less of-men than that lamb could (sg.) eat ...' (Byrhtferth's Manual: 217)

d. Swa se wyrtweart his wyrta geornor sette & plantode, swa he 'As the gardener his plants more eagerly placed and planted, so he him laes funde, ponne he eft to com, & he geseah, [thorn]aet [thorn]a of-them less found, when he back to(-them) came, and he saw that the wyrta sume waeron mid mannes fotum fortredene plants some were with of-man feet trodden-down' (Gregory's Dialogues, Bk. 1: 150)

(All the OE examples throughout, except where noted, are the product of a search of the Toronto Corpus; page nos. are those of the entry in the corpus for the item in question, here laes). In (9a) laes takes a plural partitive genitive synna (though there is much case-syncretism in this paradigm), but the number on the verb is singular. We see the same pattern with hes manna in (9b), but the verb in the comparative clause is plural. In (9e) this is reversed, with the (subjunctive) verb beon immediately preceding laes being plural while the verb in the comparative clause, maege, is singular (subjunctive). This variation is characteristic of group nouns in OE (such as folc 'people'), though, according to Mitchell, "when the collective noun and the verb are in the same clause, the verb is normally singular" (1985: 41, [section]80).

The partitive genitive construction does not survive into Modern English (ModE). And even in OE we find laes apparently not accompanied by the genitive in (10):

10) gif hit haef[eth] loes stoefgefegu, [eth]onne hit aet fruman ... haefde 'if it has less letter-combinations, than it at beginning ... had' (AElfric's Grammar: 65)

Staefgefegu bears a nominative/accusative plural ending. This illustrates the so-called "appositional" (Heltveit 1969) or "concatenative" (Heltveit 1977) construction.

The examples in (9) and (10), and the others I have inspected, are all plausibly collective, though itis difficult with respect to texts from a language not currently spoken to be certain about this. A collective interpretation would be consistent with the observed parallelism with group, or "collective", nouns with respect to number concord on the verb. If this is so, then the major change here seems to be the growth of stigmatisation of plural collectives with less at a later period.

In ModE, less, of course, does not take an inflected partitive, and is regularly accompanied by an ulmaarked noun. But ModE does have a partitive construction with quantifiers accompanied by a definite phrase:

11) some of the mud/less of the mud

whereas in OE, aecording to Heltveit, the of-construction "played an insignificant part as a partitive genitive equivalent" (1977: 78).

Let us consider now how to represent these partitive constructions, among which I include, despite the lack of overt marking, the construction in (10) and its ModE descendants in the examples with less in note 1. Let us start with the overtly partitive constructions, however.

The partitive relation can be expressed inflexionally, as in OE, or periphrastically, as in (11). Its behaviour is characteristic of a member of a functional category. We can differentiate between the periphrastic and the inflexional construction as in (12):

12) a.

[ILLUSTRATION OMITTED]

b.

[ILLUSTRATION OMITTED]

We have sets of lett-headed constructions, with each head being linked to its complements by a (solid) line which represents a dependency arc. Thus far I have left the categories of the nodes unspecified, except in the case of the partitive, which is a member of the functional category that Anderson (1997) calls functor. Functors may be realised as adpositions or inflexions: in (11a) the complement of the functor is adjoined to it, they are linearly distinct in the syntax; in (11b) the complement is subjoined, it and its head coincide linearly in the syntax. (The sequence of their realisations is assigned in the morphology). Let us now look at the other categories.

The is a determinative, a member of the class of determiners and pronouns. In terms of the system of categories of Anderson (1997), as extended in Anderson (2003), determinatives are represented as{N}: they are characterised as a primary category or word class by the sole presence of the notional feature of referentiality. Mud, as a noun, is represented as {N;P}: its representation contains both the referentiality feature and the predicability feature, since, unlike determinatives, nouns tan be predicative; but the referentiality feature is dominant over the predicability feature, indicated by its placement to the left of the semi-colon. (Verbs are {P;N}; adjeetives, as we shall sec, are characterised as having "mutual preponderance" of the two features, indicated by {P:N}). Less/laes is, whatever else, a determinative that takes a partitive functor as a complement. The functor category is eharacterised by the absence of both of these primary features, i.e. features determining primary categories (or "parts of speech"); {prt} is a secondary category associated with functors. We can thus flesh out the representations in (12) as in (13):

13) a.

[ILLUSTRATION OMITTED]

b.

[ILLUSTRATION OMITTED]

The sub-representations to the right of the slashes in (13) indicate the complement type required by the head: less/laes takes a partitive complement. {def} (definite) is, like {prt}, a secondary category, but in this case of {N} rather than {}. For simplicity, (13) doesn't inelude the information about the complementation of the functor or the article. These valencies we must examine more carefully, however.

Consider firstly the other kind of structure in which less can occur, illustrated by (14a), with a mass noun:

14) a. less mud

b. Laesse wite he [eth]rowa[eth] on helle

"less torment (acc.) he suffers in hell'

(Pastoral Care: 30)

(14b) exemplifies the OE equivalent. The use of less/laesse and the like in such examples is traditionally distinguished as "dependent" vs. the "independent" uses which have been in focus so far in this note and which also illustrated in (15), with, in this case, no overt partitive:

15) a. She won't accept less

b. [thorn]reora daga faesten, and hwilum laes, hwilom ma

'three days (gen.) fast, and sometimes less, sometimes more'

(Napier 1883: 100)

But the motivations for this "dependent"/"independent" distinction, made in such terms, so that less occurs either as head or dependent in the nominal construction, are obscure. Commonly, little argumentation has been offered in its support; and that put forward by, e.g., Selkirk (1977) is theoretically parochial. The examples in (14) are semantically partitive, and there is nothing about their syntax which is at odds with interpreting them as determinatives that take a complement, and specifically a partitive one, in all these examples, save that the construction may be elliptical, as in (15). Functional categories may be attributed to a construction without this being reflected in its morphology or periphrastically. So that we associate a locative functor with the last word in I'll see you Tuesday even in the absence of a distinct realisation for the functor, even morphologically. (14) contain implicit partitives. And we can attribute to these "concatenative" constructions the saine structure, in this respect, as in (13b). Less has changed less than one might think.

The traditional arguments against headhood for the determinative in (14) involving "optionality" are indecisive: less appears to be optional only because it necessarily takes a complement, unless the latter is ellipted. It is true that the "complement" may appear without less:

16) I want (less) mud.

But this is not uncommon: the existence of He wants a new car is not evidence that a new car is not the complement of get in He wants to get a new car. Complementation is a semantic relation that cannot be reduced to a matter of simple syntactic tests. Moreover, non-generic mud in these circumstances remains partitive: it is associated with an implicit partitive and an implieit determinative, which latter may be ruade overt as some. The structure for (non-generic) mud is then as in (17):

17)

[ILLUSTRATION OMITTED]

And we can also attribute this to the noun in (13a), whose structure we can now expand as in (18):

18)

[ILLUSTRATION OMITTED]

I have also now spelled out in (18) the complement-type required by Ne, which is a partitive-taking determinative. (This oversimplifies somewhat, since the normally takes only a non-periphrastic partitive-taker).

In OE the "concatenative" construction appears also in circumstances which now require the of-construction, as illustrated for sure (for which examples are plentiful) by (19a) (which together with the rest of (19) is drawn from Heltveit 1977: 51-4):

19)  a.
     i. sume his [eth]eawas
        'some (pl.) his customs (non-gen.)'
        (AElfric Homilies I: 115)

    ii. sume [thorn]a Denisce men
        'some the Danish men (non-gen.)'
        (Wulfstan Homilies: 350)

   iii. sume [thorn]a [eth]e [eth]aer stodon
        'some those-that there stood'
        (Anglo-Saxon Gospels: 473)

    iv. sume big woldon
        'some them (non-gen.) wanted'
        (Anglo-Saxon Gospels: 498)

     b. his apostolas sume
        'lais apostles (non-gen.) some'
        (AElfric Homilies II: 996)

     c. i. [thorn]a tel[thorn] hie brohton sume [thorn]aem cyninge
        'the teeth (non-gen.) they brought some to-the king'
        (Orosius, Bk. 1: 534)

    ii. se here ferde [thorn]a sum to Denemearce
        'the (Danish) army (non-gen.) went then some to Denmark'
        (Chronicle: 728)

     d. of [eth]am waes sum gehaten Placidus, and sum o[eth]er Maurus
        'of them was one called Placidus, and some other Maurus'
        (AElfric Homilies II: 31)

The quantifying determinative may also be postposed, as in (19b), just as when it takes an inflexional partitive (see again Heltveit 1977: 53-4). (In ModE both of the properties illustrated by (19a, b) are limited to all, both, each, and the latter is not available in all positions). Moreover, its complement may be topicalised while leaving the determinative in non-topicalised position, or otherwise separated ri'oto the quantifier, as in (19c). In ModE this too is, where at all possible, limited to the of-construction. (19d) illustrates this possibility from OE--though examples are not plentiful.

The major set of structural differences between OE and ModE in this area, then, involves the re-distribution of the inflectional, periphrastic and implicit partitive constructions and in particular the loss of the first of these and the expansion of the of-construction at the expense of both the others. The restriction on the semantics of few, that it is collective if plural, may go back to OE. I try to formulate this restriction in the next note, whose title is as much an epigraph as an illustration.

3. Swa man mare spryc[thorn] swa him laes manna gelyfe[thorn]

If count nouns are dependent on an overt determinative, the latter must be singular or plural; uncountables must depend on an overt determinative that is neither. Much in ModE is neither, and sois normally incompatible with a count noun. A(n) is singular, and sois compatible with a count noun, but not with the plural inflexion. Many and these/those are plural, and so reject uncountables, and moreover trigger presence of the plural inflexion on the noun. This/that are optionally singular, and so may take either a count or a mass noun, but not the plural inflexion. The and some are singular, plural or neither, and may accompany any noun. These familiar distributions are illustrated in (20):

20) a. Much mud, *much girl, *much girls

b. *A mud, a girl, *a girls

c. *Many/these/those mud, *many/these/those girl, many/these/those girls

d. This/that mud, this/that girl, *this/that girls

e. The/some mud, the/some girl, the/some girls

Non-generic mud on its own is associated with an implicit determinative that is neither singular nor plural. Non-generic girls on its own is associated with a plural implicit determinative.

Compare with (17), in this regard, the representation for girls in (21):

21)

[ILLUSTRATION OMITTED]

An implicit (i.e. internal, unexpressed) singular determinative is associated with singulative interpretation of a count noun, as in (22a) (vs. the non-singulative (22b)); and, unlike overt singular determinatives, this internal singulative does not inhibit the plural inflexion. Notice that if the singulative marker is made overt, the plural inflexion is lost, as shown by (22c):

22) a. Girls are eating their hearts out.

b. Girls are gathering.

c. Many a girl is eating her heart out.

If we associate (21) with the distributive reading for girls in (22a), we can represent the collective reading of (22b) with (23):

23)

[ILLUSTRATION OMITTED]

Group nouns like that in (3):

3) a. The committee is/are convening right now.

b. The committee are gnashing their teeth.

are inherently plural, and like other count nouns may take an implicit singular determinative, giving the appropriate reading for (3a).

Few and fewer are plural:

24) *Few(er) mud, *few(er) girl, few(er) girls

Little the determinative is normally neither singular nor plural:

25) Little mud, *little (? small) girl, *little (? small) girls

A few and a little are lexicalised phrases, represented as in (26):

26) a

[ILLUSTRATION OMITTED]

b.

[ILLUSTRATION OMITTED]

I do hot explore here the relationship between the paucal (pcl) feature and negation associated with e.g., few but not a few.

We have seen that while less is hot singular and is usually also not plural, it can be given a plural interpretation if it is also collective:

27) Less mud, *less girl, ?less girls

Recall (2) and (4) in note 1. This means that we can characterise less as in (28):

28) a. less = {N{pcl<pl>}/{prt}}

b.

[ILLUSTRATION OMITTED]

(28a) specifies that less is optionally ("< >") plural, and (28b) requires that, if it is, it cannot have a singulative subordinate to it, and the partitive phrase it takes is thus collective.

Plurality with OE laes may be similarly constrained in this way, as we have seen. Plural-marked complements are nevertheless more numerous than non-plurals like that in (29):

29) ... mihton [eth]a licmen laes paes sanges gehyran
    'could the body-men less of-the song hear'
    (AElfric Homilies II: 48)

(30) illustrates that OE fea, on the other hand, takes a plural complement, as does ModE few:

30) & fea monna raid heo
    'and few of-men with her'
    (Translation of Bede's Historia: 11)

The major later change here, then, is the acquisition of the stigmatisation of plural interpretation of less.

4. Comparatively little

One so-far neglected aspect of the morphosyntax of less remains for us to confront before we tan return to the connexion between the structural ambivalence of(5) and the like and the absence of a pronominal equivalent to (7b) that bears the same relationship to it as (8) does to (7a). For few and little (and many and much) are adjectives as well as determinatives: they are members of a composite category, both determinative and adjectival, so {N/{prt},P:N}. Recall the adjectival properties of (31), familiar from such discussions as Carden (1973):

31) a We have very little/*some money.

b. Our needs are few/*some.

c. The (very) little/*some money we have.

The adjectival element in the composition of these forms may be dominant, as in (31b, c), or the determinative, as in (32):

32) a. Few of our needs are pressing.

b. Little of our money remains.

Notice, moreover, that, being adjectives, few and little have comparative congeners, fewer and less, which are formed by subjoining an adjective to a comparator, and this is reflected in their morphology. Fewer and less are compared adjectives which are also, and usually primarily, determinatives, as is spelled out below in representation (36).

The comparator is another functional category, this time adjective-related, which may be expressed periphrastically (more beautiful, less beautiful) or morphologically (nicer). In the latter case, comparative-formation creates on the basis of an adjective (or adverb) a derived adjective-like category which takes a particular complement-type and which has its own characteristic modifier, or specifier, as shown in (33b) (cf. here particularly Anderson 1997: 125-128, 134-135), with (33a) instantiating the basic adjective:

33) a. Molly is (very) nice.

b. Maisie is (much) nicer than Molly.

We can represent the relevant aspects of the structure of the adjectival phrase in (33b) as in (34):

34)

[ILLUSTRATION OMITTED]

The representation for the comparator, involving ".", specifies a simple combination, without preponderance, mutual or otherwise. (Functional categories do not involve preponderances but simple combinations). I have interpreted the functor demanded by the comparative as a source, an abl(ative), but this is of no consequence here. Nor is the category of Molly--though Anderson (1997, 2003) does interpret names as determinatives.

More and less are both (morphologically irregular) lexical comparatives, as in Bill works more/less than she does, and the independent comparator in periphrastic comparisons: Bill works more/less consistently than she does. Such ambivalence is not uncommon. Would, for instance is both the irrealis form of volitional will and an independent, periphrastic irrealis marker with other verbs, as illustrated by the respective instances in If you would (only) listen to me, it would be better for everyone (Anderson 2001). In the latter role, would has taken over from the inflexional subjunctive. We find the two alternatives in the following successive apodoses from Thackeray's Henry Esmond (1852: 179, Bk. 2): Had she been a Whig, he had been one; had she followed Mr. Fox, and turned Quaker, no doubt he would have abjured ruffles and a periwig ... Irrealis in both the protases is marked by the inflexional subjunefive, as in Present-day English (Anderson 1991).

We can now turn to (5), repeated here for ease of reference:

5) She earned less/?fewer than twenty pounds.

The comparative less here takes a comparative source which contains an element which is again, whatever else it might be, a partitive-taking determinative like less, namely twenty:

35) twenty (of the) pounds

Both less and twenty take a partitive complement. Hence the ambivalence: pounds has two potential mothers. I suggest that these are actual mothers" the valency requirements of both less and twenty are satisfied by pounds; pounds complements both less and twenty. We have yet more evidence of argument sharing, which is how "raising structures" are interpreted by Anderson (1997: 230-236, [section] 3.3.4), with, for example, the pronoun in I saw him leave being dependent on both verbs.

This relationship between less and twenty is represented in (36):

36)

[ILLUSTRATION OMITTED]

Less is interpreted as a partitive-taking {N} which incorporates a comparator structure. The placement of the comparative-source phrase enables both partitive-takers to satisfy their valencies with a nominal to their right. And the fact that the plural form, here pounds, is not adjacent to the less weakens awareness of the traditional proscription against such a combination.

Argument-sharing is also involved in the overt-singulative construction of (22c):

37)

[ILLUSTRATION OMITTED]

Many is a hybrid, like few: a determinative, {N}, with an adjectival component, {N:P}. I have interpreted the singulative a as an element seeking to modify many. Such elements introduce anode above the category sought for (specified to the right of the backslash in the representation for a here), a node which has the same category as the sought-for category. This is how modifiers in general, such as adjuncts and specifiers, are characterised (Anderson 1997). The two indefinite determinatives again share their partitive. I assume, however, that this configuration is now lexicalised.

There is finally the problem of the lack of an alternative to (b) which is parallel to the relationship between (8) and (7a):

7) a. Five items or less than five (items).

b. Five or less than five items.

8) Five items or less than that.

38) a. *Five or less than that items.

b. Five or less than that number of items.

Here it seems that that cannot assume the internal position assumed by five because in these circumstances, where it does not agree in number with the following noun, it is not a partitive-taker and cannot license the occurrence of the following items. Compare (38b), which introduces a partitive-taker. The same partitive relationship (or rather the absence of it) is involved here as underlies the structural ambivalence of (5). Of course, when the comparative is part of an attributive rather than a quantifier structure, pre-position is quite normal:

39) a less than honest answer

No double motherhood is involved, in that less than honest is simply a modifier of answer rather than a partitive taking {N}, as it is in (36), and the comparison in this case involves two adjectival elements not two partitive-takers. Less, like little, is ambivalent as to quantificational versus adjectival status.

5. Conclusion

Note 1 is concerned to assert the currency of the stigmatised construction in which less is associated with plural nouns, provided these are interpreted collectively. This construction, less + plural noun, is particularly unexceptionable when it involves also a numeral as complement of a than dependent on the less. In note 2 there is an examination of the apparent OE sources of this construction. The combination of quantifier + noun is already current in OE, but also common is the construction of quantifiers with a genitive noun, an inflectionally marked partitive construction, which is absent from ModE. We also find quantifiers and nouns linked by an of, a periphrastic partitive: this is possible with a wider set of complements in OE, though the construction does not seem to be very common therein in general. The note suggests that all these combinations with quantifiers involve a partitive construction, and formulates representations of the structures involved. A further major difference between OE and Mode is the stigmatisation of the capacity of less to take plural complements. The growth of this warrants further research (or the provision of information to the present author as to where such is documented). Note 3 formulates the collective/singulative distinction, and the representation appropriate for ModE less. In note 4 it is argued that in a phrase like less than five items, items depends as a partitive on both the quantifier less and the numeral five. That here the plural item depends on less at a distance weakens the strength of the traditional shibboleth with concern for which this series of notes began.

(1) I am grateful to Fran Colman for help in the preparation of these notes, particularly note 2, and to Graeme Trousdale for his perceptive comments on an earlier version. I've already thought of the obvious sarcastic responses to the title of note 4 (as well as that of 3): original responses only, please.

(2) The abbreviations NS, ST, and TES refer to New Statesman, Sunday Times, and Times Educational Supplement respectively.

REFERENCES

PRIMARY SOURCES

Foylibra Foyle's Bookshop Magazine, June 1973.

Lyly, John 

1579

[1923] Euphues, the anatomy of wit. London: Constable.

New Statesman (NS), 17 Match 1961.

Orwell, George 

1938

[1969] Homage to Catalonia. London: Penguin.

Sunday Times (ST), 31 October 1971.

Times Educational Supplement (TES), 10 Match 1972.

Thackeray, William Makepeace 

1852

[1878] The history of Henry Esmond, Esq. London: Smith, Elder & Co.

Toronto Corpus [available from Dictionary of Old English Project, Center for Medieval Studies, University of Toronto, Canada].

SECONDARY SOURCES

Anderson, John M.

1991 "Grammaticalisation and the English modals", in: P. Kakietek (ed.), 9-27.

1997 A notional theory of syntactic categories. Cambridge: Cambridge University Press.

2001 "Modals, subjunctives and (non-)finiteness", English Language and Linguistics 5: 159-66.

2003 "On the structure of names", Folia Linguistica 37: 347-98.

Carden, Guy

1973 English quantifiers: Logical structure and linguistic variation. Tokyo: Taishukan.

Culicover, Peter--Thomas Wasow--Adrian Akmajian (eds.)

1977 Formal syntax, New York: Aeademie Press.

Fowler, H. W.

1926 A dictionary of Modern English usage. Oxford: Oxford University Press.

Gowers, Emest

1954 The complete plain words. London: H. M. Stationery Office.

[1973] [Reprinted: Harmondsworth: Penguin].

Heltveit, Trygve

1969 "The Old English appositional construction exemplified by sume his geferan: A forerunner of the modern construction a friend of mine7", English Studies 50: 225-235.

Heltveit, Trygve

1977 "Aspects of the syntax of quantifiers in Old English", Norwegian Journal of Linguistics 31: 47-94.

Jespersen, Otto

1949 A Modern English grammar, on historical principles, VII. London--Copenhagen: George Allen & Unwin and Ejnar Munksgaard.

Juul, Ame

1972 "The category of number in Modern English", English Studies 53: 1-23.

1975 On concord of number in Modern English. Copenhagen: Nova.

Kakietek, P.

1991 Problems in the modality of natural language. Opole: The Pedagogical University of Opole.

Mitchell, Bruce

1985 Old English syntax. Oxford: Oxford University Press.

Murray, J. A. H.--H. Bradley--W. A. Craigie--C. T. Onions (eds.)

1888-1933 A New English dictionary on historical principles. London: Oxford University Press.

Poutsma. H.

1914 A grammar of Late Modern English, II, Section I, A. Groningen: P. Nonrdhoff.

1916 A grammar of Late Modern English, II, Section I, B. Groningen: P. Noordhoff.

Selkirk, Elisabeth

1977 "Some remarks on noun phrase structure", in: Peter W. Culicover--Thomas Wasow--Adrian Akmajian (eds.), 285-316.

JOHN ANDERSON

Methoni, Greece